Datasets:

MMSciCode
/

MMSciCode

language stringclasses 3 values	sample_id stringlengths 18 54	func_index int64 0 2	paper_id stringlengths 3 54	paper_title stringlengths 18 201	subject stringclasses 95 values	function_name stringlengths 1 52	file_path stringlengths 8 154	paper_section stringlengths 0 225	conda_env stringlengths 0 39	docker_image stringlengths 0 50
python	001_RM-Bench_Benchmarking_Reward_Models_of_Language_Mo	0	001_RM-Bench_Benchmarking_Reward_Models_of_Language_Mo	RM-Bench: Benchmarking Reward Models of Language Models with Subtlety and Style	Evaluation Metrics	compute_accuracy	code/scripts/utils.py	Evaluation/Results - Core evaluation methodology implementation	001_rm_bench_test	mmsci-py-001-rm-bench:latest
python	001_RM-Bench_Benchmarking_Reward_Models_of_Language_Mo	1	001_RM-Bench_Benchmarking_Reward_Models_of_Language_Mo	RM-Bench: Benchmarking Reward Models of Language Models with Subtlety and Style	Evaluation Metrics	convert_robust_dataset_to_preference_dataset_list	code/scripts/utils.py	Methods/Data Processing - Dataset construction and processing methodology	001_rm_bench_test	mmsci-py-001-rm-bench:latest
python	002_TopoLM_brain-like_spatio-functional_organization_i	0	002_TopoLM_brain-like_spatio-functional_organization_i	TopoLM: brain-like spatio-functional organization in a topographic language model	Deep Learning Architecture with Spatial Organization	spatial_loss_fn	code/models/positions.py	Methods	002_topolm_test	mmsci-py-002-topolm:latest
python	002_TopoLM_brain-like_spatio-functional_organization_i	1	002_TopoLM_brain-like_spatio-functional_organization_i	TopoLM: brain-like spatio-functional organization in a topographic language model	Deep Learning Architecture with Spatial Organization	local_spatial_loss	code/models/positions.py	Methods	002_topolm_test	mmsci-py-002-topolm:latest
python	003_Knowledge_Entropy_Decay_during_Language_Model_Pret	0	003	Knowledge Entropy Decay during Language Model Pretraining Hinders New Knowledge Acquisition	Language Model Analysis and Enhancement	turn_into_entropy	code/analysis/entropy.py	Methods	003_knowledge_entropy_test	mmsci-py-003-knowledge-entropy:latest
python	003_Knowledge_Entropy_Decay_during_Language_Model_Pret	1	003	Knowledge Entropy Decay during Language Model Pretraining Hinders New Knowledge Acquisition	Language Model Analysis and Enhancement	main	code/analysis/entropy.py	Methods	003_knowledge_entropy_test	mmsci-py-003-knowledge-entropy:latest
python	003_Knowledge_Entropy_Decay_during_Language_Model_Pret	2	003	Knowledge Entropy Decay during Language Model Pretraining Hinders New Knowledge Acquisition	Language Model Analysis and Enhancement	main	code/analysis/change_parameters.py	Methods	003_knowledge_entropy_test	mmsci-py-003-knowledge-entropy:latest
python	005_Measuring_and_Enhancing_Trustworthiness_of_LLMs_in	0	005	Measuring and Enhancing Trustworthiness of LLMs in RAG through Grounded Attributions and Learning to Refuse	RAG Trustworthiness and Alignment	compute_trust_score	code/trust_eval/trust_eval/metrics.py	Methods	005_trust_align_test	mmsci-py-005-trust-align:latest
python	005_Measuring_and_Enhancing_Trustworthiness_of_LLMs_in	1	005	Measuring and Enhancing Trustworthiness of LLMs in RAG through Grounded Attributions and Learning to Refuse	RAG Trustworthiness and Alignment	compute_citation_metrics	code/trust_eval/trust_eval/metrics.py	Methods	005_trust_align_test	mmsci-py-005-trust-align:latest
python	005_Measuring_and_Enhancing_Trustworthiness_of_LLMs_in	2	005	Measuring and Enhancing Trustworthiness of LLMs in RAG through Grounded Attributions and Learning to Refuse	RAG Trustworthiness and Alignment	compute_macro_metrics	code/trust_eval/trust_eval/metrics.py	Methods	005_trust_align_test	mmsci-py-005-trust-align:latest
python	006_MAP_Multi-Human-Value_Alignment_Palette	0	006	MAP: Multi-Human-Value Alignment Palette	Multi-Human-Value Alignment	optimize_lambda	code/alignMAP/core/alignment.py	Methods	006_map_alignment_test	mmsci-py-006-map-alignment:latest
python	006_MAP_Multi-Human-Value_Alignment_Palette	1	006	MAP: Multi-Human-Value Alignment Palette	Multi-Human-Value Alignment	_dual_objective	code/alignMAP/core/alignment.py	Methods	006_map_alignment_test	mmsci-py-006-map-alignment:latest
python	006_MAP_Multi-Human-Value_Alignment_Palette	2	006	MAP: Multi-Human-Value Alignment Palette	Multi-Human-Value Alignment	sequential_optimize_lambda	code/alignMAP/core/alignment.py	Methods	006_map_alignment_test	mmsci-py-006-map-alignment:latest
python	007_Spread_Preference_Annotation_Direct_Preference_Jud	0	007	Spread Preference Annotation: Direct Preference Judgment for Efficient LLM Alignment	Direct Preference Optimization with Self-Refinement	dpo_loss	code/trl/trl/trainer/dpo_trainer.py	Methods	007_spa_test	mmsci-py-007-spa:latest
python	007_Spread_Preference_Annotation_Direct_Preference_Jud	1	007	Spread Preference Annotation: Direct Preference Judgment for Efficient LLM Alignment	Direct Preference Optimization with Self-Refinement	confidence_update	code/trl/trl/trainer/dpo_trainer.py	Methods	007_spa_test	mmsci-py-007-spa:latest
python	007_Spread_Preference_Annotation_Direct_Preference_Jud	2	007	Spread Preference Annotation: Direct Preference Judgment for Efficient LLM Alignment	Direct Preference Optimization with Self-Refinement	get_batch_loss_metrics	code/trl/trl/trainer/dpo_trainer.py	Methods	007_spa_test	mmsci-py-007-spa:latest
python	009_Brain_Bandit_A_Biologically_Grounded_Neural_Networ	0	009	Brain Bandit: A Biologically Grounded Neural Network for Efficient Control of Exploration	Biologically Grounded Neural Networks for Exploration	decide_simulation_multi_dim	code/model/Lyapunov_Worm_deconstruction.py	Methods	009_brain_bandit_test	mmsci-py-009-brain-bandit:latest
python	009_Brain_Bandit_A_Biologically_Grounded_Neural_Networ	1	009	Brain Bandit: A Biologically Grounded Neural Network for Efficient Control of Exploration	Biologically Grounded Neural Networks for Exploration	theory_calculation	code/model/Lyapunov_Worm_deconstruction.py	Methods	009_brain_bandit_test	mmsci-py-009-brain-bandit:latest
python	009_Brain_Bandit_A_Biologically_Grounded_Neural_Networ	2	009	Brain Bandit: A Biologically Grounded Neural Network for Efficient Control of Exploration	Biologically Grounded Neural Networks for Exploration	egreedy	code/MDP/agent.py	Methods	009_brain_bandit_test	mmsci-py-009-brain-bandit:latest
python	011_Standard_Gaussian_Process_is_All_You_Need_for_High	0	011	Standard Gaussian Process is All You Need for High-Dimensional Bayesian Optimization	High-Dimensional Bayesian Optimization with Standard GP	__init__	code/baselines/GP.py	Methods	011_gp_hdbo_test	mmsci-py-011-gp-hdbo:latest
python	011_Standard_Gaussian_Process_is_All_You_Need_for_High	1	011	Standard Gaussian Process is All You Need for High-Dimensional Bayesian Optimization	High-Dimensional Bayesian Optimization with Standard GP	train_model_ADAM	code/ablation/gp_ablation.py	Methods	011_gp_hdbo_test	mmsci-py-011-gp-hdbo:latest
python	011_Standard_Gaussian_Process_is_All_You_Need_for_High	2	011	Standard Gaussian Process is All You Need for High-Dimensional Bayesian Optimization	High-Dimensional Bayesian Optimization with Standard GP	BO_loop_GP	code/baselines/BO_loop.py	Methods	011_gp_hdbo_test	mmsci-py-011-gp-hdbo:latest
python	012_Oscillatory_State-Space_Models	0	012	Oscillatory State-Space Models	Oscillatory State-Space Models for Sequence Learning	apply_linoss_im	code/models/LinOSS.py	Methods	012_oscillatory_state_space_models_test	mmsci-py-012-oscillatory-state-space-models:latest
python	012_Oscillatory_State-Space_Models	1	012	Oscillatory State-Space Models	Oscillatory State-Space Models for Sequence Learning	apply_linoss_imex	code/models/LinOSS.py	Methods	012_oscillatory_state_space_models_test	mmsci-py-012-oscillatory-state-space-models:latest
python	012_Oscillatory_State-Space_Models	2	012	Oscillatory State-Space Models	Oscillatory State-Space Models for Sequence Learning	binary_operator	code/models/LinOSS.py	Methods	012_oscillatory_state_space_models_test	mmsci-py-012-oscillatory-state-space-models:latest
python	013_Attention_as_a_Hypernetwork	0	013	Attention as a Hypernetwork	Hypernetwork Attention Mechanisms	DotProductAttention.__call__	code/hyla/models/attention.py	Methods	013_attention_as_hypernetwork_test	mmsci-py-013-attention-as-hypernetwork:latest
python	013_Attention_as_a_Hypernetwork	1	013	Attention as a Hypernetwork	Hypernetwork Attention Mechanisms	MultiHeadDotProductAttention.__call__	code/hyla/models/attention.py	Methods	013_attention_as_hypernetwork_test	mmsci-py-013-attention-as-hypernetwork:latest
python	013_Attention_as_a_Hypernetwork	2	013	Attention as a Hypernetwork	Hypernetwork Attention Mechanisms	apply_fuzzy_logic	code/hyla/data/logic.py	Methods	013_attention_as_hypernetwork_test	mmsci-py-013-attention-as-hypernetwork:latest
python	014_Energy-based_Backdoor_Defense_Against_Federated_Gr	0	014	Energy-based Backdoor Defense Against Federated Graph Learning	Energy-based Backdoor Defense for Federated Graph Learning	adjust_bn_layers	code/node_code/models/GCN.py	Methods	014_energy_backdoor_defense_test	mmsci-py-014-energy-backdoor-defense:latest
python	014_Energy-based_Backdoor_Defense_Against_Federated_Gr	1	014	Energy-based Backdoor Defense Against Federated Graph Learning	Energy-based Backdoor Defense for Federated Graph Learning	select_models_based_on_energy	code/node_code/helpers/select_models_by_energy.py	Methods	014_energy_backdoor_defense_test	mmsci-py-014-energy-backdoor-defense:latest
python	015_Toward_Guidance-Free_AR_Visual_Generation_via_Cond	0	015	Toward Guidance-Free AR Visual Generation via Condition Contrastive Alignment	Guidance-Free Autoregressive Visual Generation	train_step	code/VAR_CCA_trainer.py	Methods	015_guidance_free_ar_generation_test	mmsci-py-015-guidance-free-ar-generation:latest
python	015_Toward_Guidance-Free_AR_Visual_Generation_via_Cond	1	015	Toward Guidance-Free AR Visual Generation via Condition Contrastive Alignment	Guidance-Free Autoregressive Visual Generation	__init__	code/VAR_CCA_trainer.py	Methods	015_guidance_free_ar_generation_test	mmsci-py-015-guidance-free-ar-generation:latest
python	015_Toward_Guidance-Free_AR_Visual_Generation_via_Cond	2	015	Toward Guidance-Free AR Visual Generation via Condition Contrastive Alignment	Guidance-Free Autoregressive Visual Generation	main	code/LlamaGen_finetune.py	Methods	015_guidance_free_ar_generation_test	mmsci-py-015-guidance-free-ar-generation:latest
python	016_RMP-SAM_Towards_Real-Time_Multi-Purpose_Segment_An	0	016	RMP-SAM: Towards Real-Time Multi-Purpose Segment Anything	Real-Time Multi-Purpose Segment Anything	mask_pool	code/seg/models/utils/mask_pool.py	Methods	016_rmp_sam_test	mmsci-py-016-rmp-sam:latest
python	016_RMP-SAM_Towards_Real-Time_Multi-Purpose_Segment_An	1	016	RMP-SAM: Towards Real-Time Multi-Purpose Segment Anything	Real-Time Multi-Purpose Segment Anything	forward	code/seg/models/heads/yoso_head.py	Methods	016_rmp_sam_test	mmsci-py-016-rmp-sam:latest
python	016_RMP-SAM_Towards_Real-Time_Multi-Purpose_Segment_An	2	016	RMP-SAM: Towards Real-Time Multi-Purpose Segment Anything	Real-Time Multi-Purpose Segment Anything	forward	code/seg/models/heads/rapsam_head.py	Methods	016_rmp_sam_test	mmsci-py-016-rmp-sam:latest
python	017_Residual_Deep_Gaussian_Processes_on_Manifolds	0	017	Residual Deep Gaussian Processes on Manifolds	Residual Deep Gaussian Processes on Manifolds	sphere_expmap	code/experiments/utils.py	Methods	017_test	mmsci-py-017:latest
python	018_Learning_to_Discretize_Denoising_Diffusion_ODEs	0	018	Learning to Discretize Denoising Diffusion ODEs	Learning to Discretize Denoising Diffusion ODEs	discretize_model_wrapper	code/trainer.py	Methods	018_test	mmsci-py-018:latest
python	018_Learning_to_Discretize_Denoising_Diffusion_ODEs	1	018	Learning to Discretize Denoising Diffusion ODEs	Learning to Discretize Denoising Diffusion ODEs	_train_to_match_prior	code/trainer.py	Methods	018_test	mmsci-py-018:latest
python	018_Learning_to_Discretize_Denoising_Diffusion_ODEs	2	018	Learning to Discretize Denoising Diffusion ODEs	Learning to Discretize Denoising Diffusion ODEs	marginal_lambda	code/noise_schedulers.py	Methods	018_test	mmsci-py-018:latest
python	019_Do_I_Know_This_Entity_Knowledge_Awareness_and_Hall	0	019	Do I Know This Entity? Knowledge Awareness and Hallucinations in Language Models	Knowledge Awareness and Hallucination Detection with SAE Analysis	compute_is_known	code/dataset/process_data/wikidata/check_correctness_wikidata.py	Methods	019_test	mmsci-py-019:latest
python	019_Do_I_Know_This_Entity_Knowledge_Awareness_and_Hall	1	019	Do I Know This Entity? Knowledge Awareness and Hallucinations in Language Models	Knowledge Awareness and Hallucination Detection with SAE Analysis	get_per_layer_latent_scores	code/mech_interp/feature_analysis_utils.py	Methods	019_test	mmsci-py-019:latest
python	019_Do_I_Know_This_Entity_Knowledge_Awareness_and_Hall	2	019	Do I Know This Entity? Knowledge Awareness and Hallucinations in Language Models	Knowledge Awareness and Hallucination Detection with SAE Analysis	steer_sae_latents	code/mech_interp/hooks_utils.py	Methods	019_test	mmsci-py-019:latest
python	020_TetSphere_Splatting_Representing_High-Quality_Geom	0	020	TetSphere Splatting: Representing High-Quality Geometry with Lagrangian Volumetric Meshes	TetSphere Splatting for High-Quality Geometry Representation	solve_milp	code/data/generate_init_spheres.py	Methods	020_test	mmsci-py-020:latest
python	020_TetSphere_Splatting_Representing_High-Quality_Geom	1	020	TetSphere Splatting: Representing High-Quality Geometry with Lagrangian Volumetric Meshes	TetSphere Splatting for High-Quality Geometry Representation	forward	code/energies/smooth_barrier.py	Methods	020_test	mmsci-py-020:latest
python	020_TetSphere_Splatting_Representing_High-Quality_Geom	2	020	TetSphere Splatting: Representing High-Quality Geometry with Lagrangian Volumetric Meshes	TetSphere Splatting for High-Quality Geometry Representation	forward	code/renderers/mesh_rasterizer.py	Methods	020_test	mmsci-py-020:latest
python	021_Copyright-Protected_Language_Generation_via_Adapti	0	021	Copyright-Protected Language Generation via Adaptive Model Fusion	Model Fusion Algorithm	solve_optimization	code/cp_fuse/cp_fuse/cp_model.py	Methods	021_test	mmsci-py-021:latest
python	021_Copyright-Protected_Language_Generation_via_Adapti	1	021	Copyright-Protected Language Generation via Adaptive Model Fusion	Model Fusion Algorithm	_optimize_grid	code/cp_fuse/cp_fuse/cp_model.py	Methods	021_test	mmsci-py-021:latest
python	021_Copyright-Protected_Language_Generation_via_Adapti	2	021	Copyright-Protected Language Generation via Adaptive Model Fusion	Model Fusion Algorithm	objective	code/cp_fuse/cp_fuse/cp_model.py	Methods	021_test	mmsci-py-021:latest
python	022_BIRD_A_Trustworthy_Bayesian_Inference_Framework_fo	0	022	BIRD: A Trustworthy Bayesian Inference Framework for Large Language Models	Bayesian Inference Framework	Probnetwork.forward	code/code/run/scenario_train.py	Methods	022_test	mmsci-py-022:latest
python	023_Rethinking_Reward_Modeling_in_Preference-based_Lar	0	023	Rethinking Reward Modeling in Preference-based Large Language Model Alignment	Reinforcement Learning from Human Feedback	forward_siamese	code/networks.py	Methods	023_test	mmsci-py-023:latest
python	023_Rethinking_Reward_Modeling_in_Preference-based_Lar	1	023	Rethinking Reward Modeling in Preference-based Large Language Model Alignment	Reinforcement Learning from Human Feedback	train_model	code/networks.py	Methods	023_test	mmsci-py-023:latest
python	023_Rethinking_Reward_Modeling_in_Preference-based_Lar	2	023	Rethinking Reward Modeling in Preference-based Large Language Model Alignment	Reinforcement Learning from Human Feedback	cross_prompt_data_generation	code/step5_train_rms.py	Methods	023_test	mmsci-py-023:latest
python	024_Progressive_distillation_induces_an_implicit_curri	0	024	Progressive distillation induces an implicit curriculum	Knowledge Distillation and Training Optimization	KLTrainer_progressive.training_step	code/PCFG_autoregressive/components/KL_trainer_progressive.py	Methods	024_test	mmsci-py-024:latest
python	024_Progressive_distillation_induces_an_implicit_curri	1	024	Progressive distillation induces an implicit curriculum	Knowledge Distillation and Training Optimization	KLTrainer_progressive.compute_loss	code/PCFG_autoregressive/components/KL_trainer_progressive.py	Methods	024_test	mmsci-py-024:latest
python	026_SD-LoRA_Scalable_Decoupled_Low-Rank_Adaptation_for	0	026	SD-LoRA: Scalable Decoupled Low-Rank Adaptation for Class Incremental Learning	Low-Rank Adaptation for Continual Learning	_LoRA_qkv_timm_train.forward	code/backbone/lora.py	Methods	026_test	mmsci-py-026:latest
python	026_SD-LoRA_Scalable_Decoupled_Low-Rank_Adaptation_for	1	026	SD-LoRA: Scalable Decoupled Low-Rank Adaptation for Class Incremental Learning	Low-Rank Adaptation for Continual Learning	LoRA_ViT_timm.__init__	code/backbone/lora.py	Methods	026_test	mmsci-py-026:latest
python	026_SD-LoRA_Scalable_Decoupled_Low-Rank_Adaptation_for	2	026	SD-LoRA: Scalable Decoupled Low-Rank Adaptation for Class Incremental Learning	Low-Rank Adaptation for Continual Learning	compute_ortho_loss	code/backbone/lora.py	Methods	026_test	mmsci-py-026:latest
python	027_Improving_Probabilistic_Diffusion_Models_With_Opti	0	027	Improving Probabilistic Diffusion Models With Optimal Diagonal Covariance Matching	Deep Learning - Diffusion Models with Optimal Transport	_predict_cov_x0	code/core/diffusion/dtdpm.py	Methods	027_test	mmsci-py-027:latest
python	027_Improving_Probabilistic_Diffusion_Models_With_Opti	1	027	Improving Probabilistic Diffusion Models With Optimal Diagonal Covariance Matching	Deep Learning - Diffusion Models with Optimal Transport	dt_dsdm	code/core/criterions/ddpm.py	Methods	027_test	mmsci-py-027:latest
python	027_Improving_Probabilistic_Diffusion_Models_With_Opti	2	027	Improving Probabilistic Diffusion Models With Optimal Diagonal Covariance Matching	Deep Learning - Diffusion Models with Optimal Transport	_predict_cov_prev	code/core/diffusion/dtdpm.py	Methods	027_test	mmsci-py-027:latest
python	029_MLE-bench_Evaluating_Machine_Learning_Agents_on_Ma	0	029	MLE-bench: Evaluating Machine Learning Agents on Machine Learning Engineering	Evaluation Methodology	get_familiarity_score	code/experiments/familiarity/familiarity.py	Methods	029_test	mmsci-py-029:latest
python	029_MLE-bench_Evaluating_Machine_Learning_Agents_on_Ma	1	029	MLE-bench: Evaluating Machine Learning Agents on Machine Learning Engineering	Evaluation Methodology	grade_csv	code/mlebench/grade.py	Methods	029_test	mmsci-py-029:latest
python	029_MLE-bench_Evaluating_Machine_Learning_Agents_on_Ma	2	029	MLE-bench: Evaluating Machine Learning Agents on Machine Learning Engineering	Evaluation Methodology	get_per_comp_performance	code/experiments/familiarity/familiarity.py	Methods	029_test	mmsci-py-029:latest
python	030_Subgraph_Federated_Learning_for_Local_Generalizati	0	030_Subgraph_Federated_Learning_for_Local_Generalizati	Subgraph Federated Learning for Local Generalization	Federated Learning with Graph Neural Networks	train	code/model/client.py	Methods	030_test	mmsci-py-030:latest
python	030_Subgraph_Federated_Learning_for_Local_Generalizati	1	030_Subgraph_Federated_Learning_for_Local_Generalizati	Subgraph Federated Learning for Local Generalization	Federated Learning with Graph Neural Networks	update	code/model/server.py	Methods	030_test	mmsci-py-030:latest
python	030_Subgraph_Federated_Learning_for_Local_Generalizati	2	030_Subgraph_Federated_Learning_for_Local_Generalizati	Subgraph Federated Learning for Local Generalization	Federated Learning with Graph Neural Networks	forward	code/embedder/classifier.py	Methods	030_test	mmsci-py-030:latest
python	031_Amortized_Control_of_Continuous_State_Space_Feynma	0	031_Amortized_Control_of_Continuous_State_Space_Feynma	Amortized Control of Continuous State Space Feynman-Kac Model for Irregular Time Series	Continuous Dynamical Models	parallel_compute	code/lib/sde.py	Section 3.3	031_test	mmsci-py-031:latest
python	031_Amortized_Control_of_Continuous_State_Space_Feynma	1	031_Amortized_Control_of_Continuous_State_Space_Feynma	Amortized Control of Continuous State Space Feynman-Kac Model for Irregular Time Series	Continuous Dynamical Models	get_matrix	code/lib/sde.py	Section 3.3	031_test	mmsci-py-031:latest
python	031_Amortized_Control_of_Continuous_State_Space_Feynma	2	031_Amortized_Control_of_Continuous_State_Space_Feynma	Amortized Control of Continuous State Space Feynman-Kac Model for Irregular Time Series	Continuous Dynamical Models	associative_scan	code/lib/jax_compat.py	Section 3.3 and Appendix C	031_test	mmsci-py-031:latest
python	032_Cheating_Automatic_LLM_Benchmarks_Null_Models_Achi	0	032_Cheating_Automatic_LLM_Benchmarks_Null_Models_Achi	Cheating Automatic LLM Benchmarks: Null Models Achieve High Win Rates	Adversarial Machine Learning	random_search_optimization	code/adversarial_benchmarking.py	Methods	032_test	mmsci-py-032:latest
python	032_Cheating_Automatic_LLM_Benchmarks_Null_Models_Achi	1	032_Cheating_Automatic_LLM_Benchmarks_Null_Models_Achi	Cheating Automatic LLM Benchmarks: Null Models Achieve High Win Rates	Adversarial Machine Learning	derandomize_tokens_inplace	code/adversarial_benchmarking.py	Methods	032_test	mmsci-py-032:latest
python	032_Cheating_Automatic_LLM_Benchmarks_Null_Models_Achi	2	032_Cheating_Automatic_LLM_Benchmarks_Null_Models_Achi	Cheating Automatic LLM Benchmarks: Null Models Achieve High Win Rates	Adversarial Machine Learning	create_structured_response	code/adversarial_benchmarking.py	Methods	032_test	mmsci-py-032:latest
python	033_Combatting_Dimensional_Collapse_in_LLM_Pre-Trainin	0	033_Combatting_Dimensional_Collapse_in_LLM_Pre-Trainin	Combatting Dimensional Collapse in LLM Pre-Training Data via Submodular File Selection	Data Selection Algorithm	cal_corre	code/DISF/select_disf.py	Methods	033_test	mmsci-py-033:latest
python	033_Combatting_Dimensional_Collapse_in_LLM_Pre-Trainin	1	033_Combatting_Dimensional_Collapse_in_LLM_Pre-Trainin	Combatting Dimensional Collapse in LLM Pre-Training Data via Submodular File Selection	Data Selection Algorithm	hunger_select	code/DISF/select_disf.py	Methods	033_test	mmsci-py-033:latest
python	033_Combatting_Dimensional_Collapse_in_LLM_Pre-Trainin	2	033_Combatting_Dimensional_Collapse_in_LLM_Pre-Trainin	Combatting Dimensional Collapse in LLM Pre-Training Data via Submodular File Selection	Data Selection Algorithm	cal_egin	code/Visual&verify/dominance_score.py	Results	033_test	mmsci-py-033:latest
python	036_HiRA_Parameter-Efficient_Hadamard_High-Rank_Adapta	0	036	HiRA: Parameter-Efficient Hadamard High-Rank Adaptation for Large Language Models	Parameter-Efficient Fine-Tuning	forward	code/hira/tuners/lora.py	Methods	036_test	mmsci-py-036:latest
python	036_HiRA_Parameter-Efficient_Hadamard_High-Rank_Adapta	1	036	HiRA: Parameter-Efficient Hadamard High-Rank Adaptation for Large Language Models	Parameter-Efficient Fine-Tuning	update_layer	code/hira/tuners/lora.py	Methods	036_test	mmsci-py-036:latest
python	036_HiRA_Parameter-Efficient_Hadamard_High-Rank_Adapta	2	036	HiRA: Parameter-Efficient Hadamard High-Rank Adaptation for Large Language Models	Parameter-Efficient Fine-Tuning	reset_lora_parameters	code/hira/tuners/lora.py	Methods	036_test	mmsci-py-036:latest
python	037_On_Conformal_Isometry_of_Grid_Cells_Learning_Dista	0	037	On Conformal Isometry of Grid Cells: Learning Distance-Preserving Position Embedding	Deep Learning Architecture	gridnessScore	code/source.py	Results	037_test	mmsci-py-037:latest
python	039_A_Theoretically-Principled_Sparse_Connected_and_Ri	0	039	A Theoretically-Principled Sparse, Connected, and Rigid Graph Representation of Molecules	Graph Neural Networks	project_sphere	code/mol_unit_sphere.py	Methods	039_test	mmsci-py-039:latest
python	039_A_Theoretically-Principled_Sparse_Connected_and_Ri	1	039	A Theoretically-Principled Sparse, Connected, and Rigid Graph Representation of Molecules	Graph Neural Networks	get_chull_graph	code/alignment/pyorbit/utils/qhull.py	Methods	039_test	mmsci-py-039:latest
python	039_A_Theoretically-Principled_Sparse_Connected_and_Ri	2	039	A Theoretically-Principled Sparse, Connected, and Rigid Graph Representation of Molecules	Graph Neural Networks	angle_between_vectors	code/alignment/pyorbit/utils/geometry.py	Methods	039_test	mmsci-py-039:latest
python	040_MOS_Model_Synergy_for_Test-Time_Adaptation_on_LiDA	0	040	MOS: Model Synergy for Test-Time Adaptation on LiDAR-Based 3D Object Detection	Test-Time Adaptation	mos	code/pcdet/tta_methods/mos.py	Methods	040_test	mmsci-py-040:latest
python	040_MOS_Model_Synergy_for_Test-Time_Adaptation_on_LiDA	1	040	MOS: Model Synergy for Test-Time Adaptation on LiDAR-Based 3D Object Detection	Test-Time Adaptation	aggregate_model	code/pcdet/tta_methods/mos.py	Methods	040_test	mmsci-py-040:latest
python	040_MOS_Model_Synergy_for_Test-Time_Adaptation_on_LiDA	2	040	MOS: Model Synergy for Test-Time Adaptation on LiDAR-Based 3D Object Detection	Test-Time Adaptation	hungarian_match_diff	code/pcdet/tta_methods/mos.py	Methods	040_test	mmsci-py-040:latest
python	041_Unlocking_the_Power_of_Function_Vectors_for_Charac	0	041_Unlocking_the_Power_of_Function_Vectors_for_Charac	Unlocking the Power of Function Vectors for Characterizing and Mitigating Catastrophic Forgetting in Continual Instruction Tuning	Deep Learning Architectures	compute_function_vector	code/src/fvector/utils/extract_utils.py	Methods	041_test	mmsci-py-041:latest
python	041_Unlocking_the_Power_of_Function_Vectors_for_Charac	1	041_Unlocking_the_Power_of_Function_Vectors_for_Charac	Unlocking the Power of Function Vectors for Characterizing and Mitigating Catastrophic Forgetting in Continual Instruction Tuning	Deep Learning Architectures	add_function_vector	code/src/tuning/trainer/base.py	Methods	041_test	mmsci-py-041:latest
python	041_Unlocking_the_Power_of_Function_Vectors_for_Charac	2	041_Unlocking_the_Power_of_Function_Vectors_for_Charac	Unlocking the Power of Function Vectors for Characterizing and Mitigating Catastrophic Forgetting in Continual Instruction Tuning	Deep Learning Architectures	training_step	code/src/tuning/trainer/base.py	Methods	041_test	mmsci-py-041:latest
python	042_Robustness_Inspired_Graph_Backdoor_Defense	0	042_Robustness_Inspired_Graph_Backdoor_Defense	Robustness Inspired Graph Backdoor Defense	Graph Neural Networks and Backdoor Defense	sample_noise_all	code/defense.py	Methods	042_test	mmsci-py-042:latest
python	042_Robustness_Inspired_Graph_Backdoor_Defense	1	042_Robustness_Inspired_Graph_Backdoor_Defense	Robustness Inspired Graph Backdoor Defense	Graph Neural Networks and Backdoor Defense	prediction_variance_calculation	code/defense.py	Methods	042_test	mmsci-py-042:latest
python	042_Robustness_Inspired_Graph_Backdoor_Defense	2	042_Robustness_Inspired_Graph_Backdoor_Defense	Robustness Inspired Graph Backdoor Defense	Graph Neural Networks and Backdoor Defense	fientune	code/models/GCN.py	Methods	042_test	mmsci-py-042:latest
python	043_Proxy_Denoising_for_Source-Free_Domain_Adaptation	0	043	Proxy Denoising for Source-Free Domain Adaptation	Source-Free Domain Adaptation	train_target	code/src/methods/oh/ProDe.py	Methods	043_test	mmsci-py-043:latest
python	043_Proxy_Denoising_for_Source-Free_Domain_Adaptation	1	043	Proxy Denoising for Source-Free Domain Adaptation	Source-Free Domain Adaptation	test_time_adapt_eval	code/src/methods/oh/ProDe.py	Methods	043_test	mmsci-py-043:latest
python	043_Proxy_Denoising_for_Source-Free_Domain_Adaptation	2	043	Proxy Denoising for Source-Free Domain Adaptation	Source-Free Domain Adaptation	IID_loss	code/src/utils/IID_losses.py	Methods	043_test	mmsci-py-043:latest
python	044_On_the_Identification_of_Temporal_Causal_Represent	0	044	On the Identification of Temporal Causal Representation with Instantaneous Dependence	Temporal Causal Representation Learning	NPInstantaneousTransitionPrior.forward	code/IDOL_synthetic/IDOL/modules/components/transition.py	Methods	044_test	mmsci-py-044:latest
python	044_On_the_Identification_of_Temporal_Causal_Represent	1	044	On the Identification of Temporal Causal Representation with Instantaneous Dependence	Temporal Causal Representation Learning	Model.loss_function	code/realworld/model/IDOL.py	Methods	044_test	mmsci-py-044:latest
python	044_On_the_Identification_of_Temporal_Causal_Represent	2	044	On the Identification of Temporal Causal Representation with Instantaneous Dependence	Temporal Causal Representation Learning	InstantaneousProcess.loss_function	code/IDOL_synthetic/IDOL/modules/instantaneous.py	Methods	044_test	mmsci-py-044:latest
python	046_Learning_Dynamics_of_LLM_Finetuning	0	046	Learning Dynamics of LLM Finetuning	Deep Learning Optimization and Training Analysis	preference_loss	code/src/trainers.py	Methods	046_test	mmsci-py-046:latest
python	046_Learning_Dynamics_of_LLM_Finetuning	1	046	Learning Dynamics of LLM Finetuning	Deep Learning Optimization and Training Analysis	_get_batch_logps	code/src/trainers.py	Methods	046_test	mmsci-py-046:latest

End of preview. Expand in Data Studio

MMSciCode

About | Benchmark Construction | Statistics | Usage | Citation

About

This repository contains MMSciCode, a benchmark for paper-grounded scientific research coding. MMSciCode evaluates whether a model can recover masked core functions from real research code using the surrounding repository context, paper-derived context, and sample-specific implementation metadata.

The benchmark spans Python, R, and C/C++ projects collected from scientific papers and their associated code releases. Each task is evaluated by inserting the generated function back into the original project and running the corresponding unit tests.

This Hugging Face dataset repository contains both the benchmark data and the Dockerfile build assets used to reproduce the execution environments.

Benchmark Construction

MMSciCode is built from real scientific software projects through a function-level construction pipeline:

Paper and code collection: scientific papers are paired with their released code repositories.
Function extraction: candidate functions are extracted from each project together with file paths, line numbers, and repository structure metadata.
Core-function selection: expert annotations identify functions that implement paper-relevant algorithms, equations, simulations, or analysis procedures.
Task creation: selected functions are masked while preserving the surrounding code context and paper-derived implementation evidence.
Executable validation: generated implementations are inserted back into the original project and checked with sample-specific tests.
Environment packaging: Dockerfile build contexts are provided for the exported execution environments.

Why MMSciCode?

Scientific coding differs from short standalone programming tasks: solutions must often match paper-specific notation, domain assumptions, project-local APIs, numerical behavior, and existing repository structure. MMSciCode is designed to measure those abilities directly by using real research code and containerized execution.

Each sample directory under a language-level data/ folder includes metadata such as available functions, selected core functions, paper or article context, repository structure, and test status.

Statistics

Item	Count
Function-level tasks	624
Source sample directories	285
Programming languages	3
Python samples	203
R samples	60
C/C++ samples	22
Dockerfile environment directories	204

Dockerfile environment directories are organized by language-level dockerfiles/ folders:

Dockerfile group	Count
`Python/dockerfiles/`	201
`R/dockerfiles/`	1
`C_CPP/dockerfiles/`	2

Repository Layout

MMSciCode/
  Python/
    data/
      <sample_id>/
    dockerfiles/
      <environment_id>/
  R/
    data/
      <sample_id>/
    dockerfiles/
      <environment_id>/
  C_CPP/
    data/
      <sample_id>/
    dockerfiles/
      <environment_id>/
  manifest.jsonl
  index.tsv
  build_all_serial.sh
  distributable_env_dockerfiles.tar.gz

Each sample directory contains the following files. Required files are present in every sample; optional files are present when applicable.

File	Status	Description
`selected_core_functions.json`	required	The functions selected for evaluation: `function_name`, sample-relative `file_path`, description, paper reference, formula, key-term mapping, and implementation cues.
`unit_test_status.json`	required	Execution environment (`environment.conda_env_name`) and the per-function test wiring (`target_functions[].src_file` / `reference_file`). See notes below.
`article_content.json`	optional	Parsed paper text (abstract / sections) used to build paper-grounded prompts.
`article_info.json` or `article_metadata.json`	optional	Paper title, URL, and subject.
`functions.json`	optional	Full inventory of functions extracted from the project (informational; not required by the evaluation pipeline).
`structure.txt`	optional	Repository directory tree of the original project.
`code/` or the project root dir	required	The original project source the masked function is drawn from.

Field-level notes for unit_test_status.json:

target_functions[].line_start / line_end are optional and may be null; they are informational and are not consumed by the evaluation pipeline (function location is resolved from selected_core_functions.json).
target_functions[].test_file is optional and may be empty for samples whose harness discovers tests by convention.
legacy_backup and validation hold historical build/validation records and are not required to run the benchmark.

The root files:

manifest.jsonl — one row per benchmark task (624 rows) with language, sample_id, func_index, paper_id, paper_title, subject, function_name, file_path, paper_section, conda_env, and docker_image. A task is uniquely identified by (sample_id, func_index), where func_index is the 0-based position in that sample's selected_core_functions.json. file_path is the sample-relative path of the file the function lives in and is tested in. This is also the file rendered by the Dataset Viewer.
index.tsv — maps each Docker image name to its environment id and Dockerfile directory (image, env, dir, editable). The editable flag is Docker build metadata (whether the environment installs the project as an editable package); it does not indicate whether a task may be modified.
distributable_env_dockerfiles.tar.gz — the same Dockerfile assets as a standalone package.

Note on original project code. Each sample bundles a real research project. That upstream source may contain the original authors' absolute paths, machine-specific comments, or non-English text. These are part of the preserved research artifact and are intentionally left unmodified; the MMSciCode-generated metadata files above have been scrubbed of any build-time paths.

Note on standalone C/C++ samples. A small number of C/C++ samples compile with the system toolchain and have no conda_env_name (and therefore no Docker image / empty docker_image in the manifest). They are evaluated with a local gcc/g++ + cmake build rather than a prebuilt container.

Usage

Downloading the Dataset

git lfs install
git clone https://huggingface.co/datasets/MMSciCode/MMSciCode
cd MMSciCode

Or download with huggingface_hub:

from huggingface_hub import snapshot_download

dataset_dir = snapshot_download(
    repo_id="MMSciCode/MMSciCode",
    repo_type="dataset",
)

Inspecting a Task

Each benchmark task is defined inside a language-specific data/ directory. For example:

ls Python/data/<sample_id>
cat Python/data/<sample_id>/selected_core_functions.json
cat Python/data/<sample_id>/unit_test_status.json

selected_core_functions.json describes the functions selected for evaluation, including their source locations, natural-language descriptions, paper references, and implementation cues.

Building Docker Environments

The repository includes Dockerfile build contexts under the language-level dockerfiles/ directories. To build all indexed environments serially:

chmod +x build_all_serial.sh
./build_all_serial.sh

The build script reads index.tsv, locates each Dockerfile directory under Python/dockerfiles/, R/dockerfiles/, or C_CPP/dockerfiles/, and tags the resulting image with the name listed in the index.

Optional build arguments can be passed through environment variables:

CONDA_MIRROR=https://mirrors.tuna.tsinghua.edu.cn/anaconda ./build_all_serial.sh
PIP_STRICT=1 ./build_all_serial.sh

Using the Standalone Dockerfile Package

If you only need the Dockerfile build contexts, extract the bundled archive:

tar -xzf distributable_env_dockerfiles.tar.gz
cd distributable_env_dockerfiles
./build_all_serial.sh

Links

Resource	Link
Dataset	MMSciCode/MMSciCode
Organization	MMSciCode
Paper	ACL 2026
Code	github.com/MMSciCode/MMSciCode
Dockerfile index	`index.tsv`
Docker build script	`build_all_serial.sh`

Citation

If you find MMSciCode useful in your research, please cite our paper:

@inproceedings{xia-etal-2026-mmscicode,
    title = "{MMS}ci{C}ode: Real-world Evaluation of Multilingual Multi-Discipline Scientific Research Coding",
    author = "Xia, Xue and Yang, Zheyuan and Cohan, Arman and Zhao, Yilun",
    editor = "Liakata, Maria and Moreira, Viviane P. and Zhang, Jiajun and Jurgens, David",
    booktitle = "Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
    month = jul,
    year = "2026",
    address = "San Diego, California, United States",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2026.acl-long.1566/",
    doi = "10.18653/v1/2026.acl-long.1566",
    pages = "33981--33999",
    ISBN = "979-8-89176-390-6"
}

Downloads last month: 3,082